Skip to content

Treasury Dragons - Cash Forecasting - 2025-12-09

Metadata

  • Date: 2025-12-09
  • Event: The Treasury Dragons vs Cash Forecasting - 2025 Episode 2
  • Type: Industry Event / Conference
  • Host: Mike (Treasury Dragons)
  • Dragons (Expert Panel):
  • Varan (Zanders - 13 years treasury consulting)
  • Royston Decoster (Ferguson - Assistant Treasurer, 37 years treasury experience)
  • Andy Gifford
  • Kate (Protiviti - business development, former international banking)
  • Vendor Presenters:
  • John Paquette (TIS - CPO)
  • Matthias Van Camp (Cobase)
  • Adi Barak (Panax - VP Product)
  • Gurjit Panu (Palm - Co-founder/CEO)
  • Palm Participants: Gurjit Panu
  • Domain Areas: Cash Forecasting, Variance Analysis, Scenario Planning

Summary

Context

Annual industry webinar where treasury technology vendors ("pitchers") present cash forecasting solutions to a panel of treasury experts ("dragons"). Four vendors presented: TIS, Cobase, Panax, and Palm. Expert panel provided questions and feedback after each presentation.

Audience Poll Results

  • 78% of audience: "It's a good guide, but we could do better" (medium-term forecast accuracy)
  • 8%: "Unhappy - way out most of the time"
  • 4%: "Very happy - usually very close"
  • 7%: "What cash forecast?" (no forecasting in place)

Key Discussion Points

TIS (John Paquette): - Focus on data integration complexity - connecting ERPs, FP&A systems, TMS, bank statements - "Walk, run, fly" implementation model for gradual adoption - Working Capital Insights module tying DSO/DPO to cash forecasting - AI for analyzing customer payment behaviors and forecasting collection patterns - Emphasis on collaboration and two-way communication with entities

Cobase (Matthias Van Camp): - Strength in bank connectivity (origin as multibank payment platform) - Workflow/approval flows for central treasury control over entity inputs - AI for transaction categorization - Challenge of factoring/invoice discounting arrangements acknowledged

Panax (Adi Barak): - "AI native" positioning - AI in every layer from infrastructure to UX - Target: Mid-market companies (complex needs, lean teams) - LLM-powered chat experience for insights - 3-8 weeks typical implementation time - Emphasis on cleaning fragmented data from multiple sources

Palm (Gurjit Panu): - AI native cash flow forecasting that sits on top of TMS/ERP - Case study: ON achieved $350M idle cash unlocked - "Intelligent upload" feature - drop files as-is, AI maps columns to categories - Emphasis on confidence + accuracy (not just accuracy alone) - Assumption layering for scenario planning - Granular variance analysis (category-level across all accounts)

Dragon Panel Insights

Varan (Zanders): - Cash forecasting remains most challenging process for treasury - Key asks: What-if scenarios, seasonality handling, move away from spreadsheets - Companies with fragmented tech, multiple banks, multi-geography struggle most

Royston Decoster (Ferguson): - 37 years experience, currently 13 cloud-based treasury solutions - "Shortened forecasts becoming more reliable, but long-term still depends on human judgment" - Key challenges: Data quality inconsistent, subjective assumptions, user adoption when tools difficult - Still using spreadsheets at Ferguson despite trying alternatives

Andy Gifford: - Impressed by collaboration focus (TIS) - Valued "confidence in forecast" concept (Palm) - Noted tools are maturing rapidly

Kate (Protiviti): - "Walk, run, fly" implementation approach important for unready clients - Change management is major challenge, especially for global multinationals

Pain Points Mentioned

  • Data fragmentation - Banks, ERPs, investment platforms, payment processors all separate
  • IT dependency - Treasury teams relying on IT/BI teams to clean and prepare data
  • Spreadsheet default - Even companies who've tried tools revert to Excel
  • Entity collaboration - Getting subsidiaries to provide timely, accurate forecasts
  • User adoption - Complex tools slow adoption
  • Data quality - Inconsistent across sources
  • Forecasting relies on subjective assumptions - Long-term especially dependent on judgment

Competitive Intelligence

Vendor Positioning Differentiator
TIS Enterprise multinationals Strong data integration, entity collaboration workflows
Cobase Bank connectivity first Native multibank platform, treasury workflows
Panax AI native for mid-market Fast implementation (3-8 weeks), lean team focus
Palm AI native overlay Sits on top of TMS/ERP, confidence + accuracy emphasis

Notable Quotes

"Cash forecasting remains one of the most challenging processes for Treasury... one of the most complex process to have a reliable cash forecast." - Varan (Zanders)

"It doesn't matter how accurate your forecast is if you're not confident or trust in how it was generated, then it's a bit useless." - Gurjit Panu (Palm)

"Data quality is still often inconsistent. Forecasting relies on subjective business assumptions, and user adoption slows when the tools are difficult to use." - Royston Decoster (Ferguson)

"AI is there to help us as a tool. So we should not discount it, but do not trust it implicitly." - Royston Decoster (Ferguson)

"The biggest pain is in gathering all the data into one platform." - Matthias Van Camp (Cobase)

Action Items

  • [ ] Review Palm webinar recording from January 21st mentioned by Gurjit
  • [ ] Note competitive positioning for sales enablement

Full Transcript

Introduction & Dragon Introductions

Them: Thank you, Mike, for the introduction. Pleased to be here today on the panel. For everyone in this call. My background is in treasury consulting for the last 13 years and I work for Zonders, as Mike mentioned. And for those of you who are not familiar with the company, we are a tragic end risk management consulting firm which is headquartered in Netherlands and offices in 12 different countries across the globe. So typically we work with corporate clients and PE firms to advise them on treasury strategy and technology. And through this work, I have observed, Mike, that cash forecasting remains one of the most challenging processes for Treasury. As Andy mentioned earlier, that it has evolved through the time, but still it is one of the most complex process to have a reliable cash forecast. And especially for those companies with a large, fragmented technology landscape, multiple banking partners and operations across different geographies. So what I'm looking forward is seeing tech solutions that tackles these challenges for the treasury team to produce reliable cash forecast on a centralized platform. And enabling them to perform various kind of analysis what if scenarios and the flexibility to incorporate seasonality trends in their forecast directly in the system. So that the corporates, they can move away from the mail processes. Thank you. So it's goodbye spreadsheets as far as you're concerned, Varan. And of course, those seasonal variances. And that idea of bringing in multiple entities across the globe. Really important. Thank you. Varan Royston Decoster, A familiar face to many of us, but for those who don't know you just let everyone know who you are and what you're looking for today. Thank you, Mike. Yes. Welcome, everyone, to this call. Well, penultimate call, the assistant treasurer, folks in North American, leading value added distributor of plumbing and heating products. Turnover is just over $31 billion. And we're prime listed on New York Stock Exchange. My responsibilities have been really looking after the intercompany loans and the treasury technology for our group, as well as supporting the US treasury team on large ad hoc transactions. Been Trudy over 37 years. 23 of those years we've been with Ferguson, worked for companies like Sky, Gillette, Polygram, Seagram and Vendi Universal Fairly wide international blue chip companies. Currently we have 13 cloud based solutions which include Cooper and Oracle as our ERP Cash flow forecasting Right interesting and well topic that most graduates have in their top three list. No different price of folks. And my view, cash forecasting. The market's definitely modernizing quickly but unevenly. No surprises. Many firms still using spreadsheets, including ourselves. Not for lack of trying because with various other challenges we'll probably see and talk about through the session, but I think mostly in Treasuries are moving towards an integrated TMS and ERP platform or cloud based system. And of course Bank APIs are there for real time data. AI machine learning are becoming without question the standard, improving accuracy, spotting anomalies and making forecasting more proactive. Shortened forecasts are becoming more reliable thanks to again real time feed, but long term forecasts still depend heavily on human judgment and scenario planning. Advanced tools now support this multi scenario modeling, stress testing, multi entity consolidation and multi currency forecasting. Obviously important for global liquidity management. Treasures though increasingly I believe one real time visibility support their working capital and liquidity optimization. And of course RegStream needs TMS provides. I'm betting some of this forecasting capabilities or partnering specialist vendors. While API driven ecosystems are helping to reduce these errors and streamline the data collection. So despite technological progress, key challenges remain. The data quality is still often inconsistent. Forecasting relies on subjective business assumptions, and user adoption slows when the tools are difficult to use. So very quickly ending with looking ahead. My view priorities include forecasts using forecasting to improve working capital, enhancing ESG related liquidity planning. And strengthening cybersecurity. As these strategic processes become more real time connected. So on that note, Mike, I'm going to look forward to what the E3.4 presenters have today. And over to you. Thank you very much, Royston. Adding still more elements to what we hope to see from our presenters. Not much to live up to here. So, Kate, again, a familiar name to many of our attendees today, but remind us who you are. Give us a sense of what you hope to see. Okay. Thanks so much, Mike. Okay, so, as you say, the business development sales for protective group. I have a long and rich background in commercial banking, working for five different international banks. So seeing it, if you will, from the other side being why Binam Road, JPMorgan City and ING, and I was very involved in supporting clients. Over the years trying to look at their cash and liquidity management positions, and, of course, also as a consultant, looking at the different providers and their systems. So I think both of my colleagues have very well outlined what we're looking for. I always have my eye to innovation and to see what's new and different. And particularly effective for clients in the future. Thanks, Meg. Thank you, Kate. So we've seen what our dragons are looking for. So who's likely to meet those requirements? The best that can tell us today tis Cobase, Panax and Palm. I'm going to turn to the first of those in just a moment. Before I do, let's check in. With our poll and see what our audience today is thinking. Well, 78% of us have chosen the middle option. I asked how happy you are with the accuracy of your medium term cash forecast. The middle answer in the middle. It's a good guide, but we could do better. 79% now of our audience are selecting that fully 8% are unhappy. It's way out most of the time. Well, we're here to help. So maybe by the end of this session you might have some ideas on how to make it better. And 4% are very happy. It's usually very close. I'm not even going to suggest. That you might be deluding yourself with that responses, but if that is true, then congratulations. You're the very small group of elite treasurers who've managed that. And congratulations to the 7% who selected what Cash Forecast. Yes, we love honesty on Treasury Dragon, so 74% right in the middle and could do with some extra help. And let's just remind ourselves of what's happening out there in the wider world beyond cash. Forecasting our prize today for lowest temperature so far comes from Ramon in Quebec, who is at minus 15 in Canada. I can see that. Jennifer in Connecticut reminds us it's sunny, but frigid. And so does Bruce. Hello, Bruce. Lynn in Connecticut and most people getting rainy in the Netherlands. Mark de Wilde says it's rainy. And Kodak from New Jersey says it's sunny and cold and Stockholm cloudy again. Thank you. Emma from Stockholm and Matthias from Amsterdam and also sunny and cold in Chicago. When is it? Not. Thank you, Chris Moses. If you haven't checked in yet, do one small plea. Rodrigo from Sao Paulo in Brazil. Cheer us up. Tell us, please, that it's sunny and warm in Brazil. It quite often is. That would make the rest of us a little happier about our midwinter bleakness. So that's enough chat. We've got a problem to solve. Everyone would like a better cash forecast and the first person to help us do that, Dragons. Thank you very much. I'll see you very soon for your responses. While I introduced our first presenter of the day, who is, of course, John Paquette. CPO Tis John by now you count as a veteran, so you know exactly what you're doing. I will tell you now that you're seven minutes are on the egg timer of doom, which is just about visible, is starting now, and it's over to you. All right. Thanks a lot, Mike. And great to be back for the last session of the year here on Treasury Dragons. Maybe you're starting here with a short research piece from an excerpt we published with Strategic Treasurer back in May. Obviously, the expectations around cash forecasting are increasing. They have been increasing for a number of different years. And what are really people looking for from a management perspective within their cash forecast? They want greater accuracy, they want more frequency of the forecast, they want more scenario building capabilities, they really want to be able to understand variances, why they're occurring. And they really want to know the opportunities for AI to both smooth out that forecast, increase the accuracy, but also help them drill into sort of really the deep insights there. So no real surprises here. But what this really says to me is that what people really want is a forecast that's more than just a report that's published out periodically, monthly. What they want is really a blueprint for how they manage liquidity. And they want to be able to do that confidently and they want to have the capabilities in there to be able to see around corners, understand what the risks are, and be ahead of all these different risks. Right. I think tis we really understand this fundamentally. We've been tailoring our product roadmap towards these changing expectations for a number of different years. And I think we really have a unique vantage point on the topic as well, where we primarily work with, say, large multinational corporates, those that have a lot of geographic complexity, entity complexity, bank complexity, a number of different sources. And for them, really, this complexity gets in the way of them really building a forecast that can really be used to drive strategy. And this is something that we fundamentally help with a lot in our product, I think makes us a bit unique as well. How do we do this? So one type of complexity that I say we saw particularly well is really the technical complexity of integrating the data necessary for cash forecasting. And to do this, we really lean on our strong data integration capabilities. And getting this piece correct is absolutely critical for foundational, for layering anything on top of this, like AI or advanced forecasting capabilities or scenario planning, making sure you have the right data in place to drive the forecast. And this is sort of step one for us with the customer, is that we'll sit down, we'll actually map out all the cash flows they want to forecast. To the right source for each one of those cash flows. Right. What's going to really drive the most accurate forecasts and the best insights for our customers? So maybe some good examples of this is a lot of our customers will want to save for their payables, receivables, information, tap into their invoice level AP and AR data within the ERP system. For payroll. It's a little bit more predictable. They might look for historical trends from their bank statement. They might go their TMS for things like deal maturities and then the FPA systems for, say, longer range expense and revenue items that they want to project out. And they certainly don't have to, you know, go to that level of sophistication off the bat. They're able to sort of grow. With us and use more and more sophisticated forecast methodologies over time. And then we use turnkey connectors to really eliminate the IT burdens of integrating this data, which is historically has helped a lot of companies back. Right. And the results from this is obviously a forecast that's going to be more accurate if it's derived from the best forecast information, the best indicators of cash flows within the organization, it's going to result in much better reporting because you're tapping into very, very good granular data and particularly variance analysis to really understand the causes of those variances. And then also, once you've integrated this data in, you can forecast as frequently as you'd like, right? So it's just about setting the scheduler up to pull this data in based on your schedule here. And you can, you know, really automate a huge portion of the forecast as you're, as you're creating the different periods. But for us, this is really just one piece of the equation. You know, the other piece of the side of the complexity for corporates that we really solve well is, you know, I call it the organizational complexity. And this is really the reliance that companies have on multiple different entities to really bring in forecast inputs. Most of our customers have, say, 30, 50, two, 50, a thousand different entities across the world, and they want to involve them in the forecasting process to really bring in those qualitative inputs. They need them to, say, confirm certain data or even provide certain data inputs. And then likewise, they want to be able to understand the local cash trends. You know, what's happening in that region or within a particular cash pool. But they also want to be able to roll that up to a more, say, high level forecast, all the way up to a corporate level forecast to understand those trends both locally and how they impact the corporate at higher levels there. And then likewise, they want it to be a two way communication where they're not just receiving inputs from their entities, but also providing them with some feedback that they're getting from the forecast and some insights that help them run things locally. And this is something that you'll see infused across our entire platform as well as you know, workflows, tasks alerts, comments and capabilities, report exports. Really everything that a business needs to be able to collaborate across the entire business, create one harmonized forecast that both works at the corporate level and down to the local level. Right. So something that we put a very strong focus on. And once you have these two, you know, really foundational aspects of creating the cash forecast in place, that's when you're in a really good place to be able to layer on, say, more advanced technologies. That can help you both improve the forecast as well as drive actual insights. And, you know, some of those are sort of referenced on this, this page here. Of course, AI, you know, to improve forecast accuracy. The way we think about AI is that, you know, with the type of companies that we work with, They're buried in mountains of data, right. And they want to be able to understand, you know, sort of what insights are trapped within all that data that aren't available to the human eye. Where are the risks in the forecast where we could potentially be off? Where are the opportunities? And this is where we're really sort of implementing AI with the radar product. We also have a Working Capital Insights module that ties both the cash forecasts into the working capital. Aspects. To help companies understand things like DSO and dpo cash conversion cycle. You know, really be understand how cash and working capital tied together and where the opportunity might be to improve on either side. And you can even use this sort of in your variance analysis. Say, if you notice that your cash receipts forecasts is off this week. And, you know, is it due to, say, a rising DSO in one region or one entity? And how do I really drive into that and see sort of what is really sort of going wrong within that entity? And how do I ultimately collaborate across the business to. To improve it? Right. So, and then scenario planning, we offer our customers a number of different ways to create different scenarios. This is really a key thing that I think most corporates are looking for these days. And this can be both building scenarios on top of, you know, the, the forecast data that we're bringing in, using different, say, forecast methodologies and things like that, but also creating more turnkey scenarios using things like the working capital data or the AI data to be able to, you know, sort of see around corners and build some quick scenarios that way. And then really be able to dive deep into the data. You know, since we're ingesting again, the most granular data, the best sources for the forecast within the business, we can provide deep levels of reporting as well that help companies really pinpoint, say, trends within there variances and things like that. And really understand them at a very, very deep level. Right. So. And then, you know, like. Like I mentioned before, it's more. It's not about creating report, it's about really creating business value for our customers. So it's all about the benefits that they can achieve. And this is going to differ a little bit from customer to customer, obviously. And this is where we really center the forecasting process is in terms of the objectives that they're looking to meet and how we can help them meet them. But there certainly are a couple here that are standardized that we see kind of consistently across most of our customer situations. You know, they're looking to save time. They don't want their, you know, sort of scarce treasury resources spending a lot of time normalizing data, collecting inputs, aggregating data to get together with this, you know, to get this global forecast together. They want that to be automated. And, you know, with that, obviously, comes the ability to be able to run the forecast more frequently and get the benefits of, you know, sort of that more frequent forecasting gives you, in terms of, you know, cash management capabilities and things like that. They're often obviously looking for proactive cash savings. They're looking to understand where the trends are that allow them to invest excess cash, pay down debt, repatriate cash, cut a dividend, whatever it might be. So we're bringing up those insights within the forecast. And they're looking for the tie ins to working capital. They want to know how to improve DSO DPO cash conversion cycle. They want to use the platform as like a real collaboration tool to, you know, sort of show the insights that treasury is seeing and work collectively with the AR and AP departments, ultimately improve working capital. And liquidity for everybody. So, you know, some of the common things that we're seeing here, and maybe just to wrap up here, a couple of great, you know, success stories, examples of this that you guys can check out on our website if you'd like to Catalogs is a company, a customer of ours, that achieved a tremendous cash benefit from implementing the tool. Unilever, you know, typical sort of multi entity global business that gained a lot of the benefits from, you know, sort of our workflows and collaborative aspects of our platform. So just kind of a couple different scenarios I thought you guys might be interested in checking out. After the President. We. We are up to time slightly over, which is unusual for you, actually. Generally, you are the absolute Timekeeper. But not a problem we've got there, and we've got a pretty good idea of what the platform can do. So it's now time to move forward and have a look at questions. I can see we have some beginning to come in from the audience, but if you have a question, type it in down there, bottom right in the question section. I will get to it if I can. Before we get there, Kate Pole. Can I turn? To you first for a question to John and tis. Absolutely. Hi, John. Great presentation as always, and very enjoyable. You know, I hear about what you're doing and how you're doing it, but. And I love the phrase blueprint for managing the liquidity, but we all know that some of these pesky clients. I say this with tongue in cheek, of course, aren't always ready, aren't always prepared. They might not have the discipline to really use the system optimally. So how do you help them get there? What do you see as the key obstacles, and how are you helping them overcome it? Yeah, I think it's a common thing we see with our customers as well. So we've kind of implemented an approach to, say, implementation that we call the walk, run, fly model, where, you know, you can get a very basic forecast off the ground using maybe the data that the company has available today. Really targeted at what the say, what's really holding them back Is that the aggregation of the data globally? Is it the, you know, say, pulling in the bank information or getting the integrations with the ERPs, or the variance analysis, focusing on that first and then really leaving them open to be able to then move to more advanced methodologies down the road, like integrating more Source System data, FP&A data for mid to long range forecasts. Obviously the AI and working capital capabilities. But so we typically go about it that way. And the other thing that we try to do is really put in place a process that's going to be adopted by the local entities, because change management could obviously be a huge problem with cash forecasting, particularly for global multinationals, where you need to get the entities on board with the process to really be able to derive the benefits. So something that's simple, collaborative and, you know, I mentioned sort of the two way exchange. Of information so that the local entities do see the benefit of the forecasting system and don't just see themselves as a giver of data, but also sort of a provider or a recipient of the insights. Thank you. Excellent. Thanks, Kate. Mark, I can see you have a question. Other dragons. I'll come to you in just a moment when I've dealt with this question from Mark. Raise your hands if you want to join in. Sir John, Mark asks, how much do you make use of AI to complete gaps in the data? Interesting. And? Create a better fog. You did mention machine learning and other things, but is AI at the point yet where you've got a gap? Can you sort of bridge that gap by putting in what feels right to an AI input? Yeah, I think it can. I mean, some of the most, I guess the ways that our customers have seen the most benefit of it are definitely on the, say, inbound customer receipts side, which are historically very, very difficult to predict. So they might be, say, baseline projecting those receipts based On I have 30 day payment terms with this customer, and I'm expecting cash in 30 days. But we can actually build on top of that logic with both our Working Capital Insights module, which can tell them, say, historical sort of payment behavior of that customer. As well as AI that can get really deep into those insights and tell them things like, hey, you might have this projected out in 30 days, but we can see based on the historical patterns, that this particular customer only makes an AP run once a month, on the third week of the month, and you're not going to want to project out the receipt. Until then, in this way, we're kind of, like I said, helping our customers see the risks within the forecast, within that mountain of data, really one of the pieces that could go wrong within the forecast that they should be paying attention to and maybe adjusting for, or at least building another scenario. Around. Makes absolute sense. Thank you for that. Royston. I think you may have a subsidiary question. Just to remind you on mute. Thank you. Yeah. Thanks, John. That great presentation again. So you've got a surprising. I'm going to ask you. The question I'm going to ask now is basically we're talking about AI, admittedly, but I'm just curious, how much attention are you paying to, for example, agentic AI in the way you're processing the data. 247 kind of potential. And also alongside that, are you thinking of perhaps implementing some sort of GPT chatgpt type? Functionality in your product in the future. Yeah, so we actually have, in a couple different respects, within the forecasting platform, we've introduced sort of an agentic AI, say GPT type functionality to be able to configure the platform more easily, which is helping our customers see quicker time to value things like how do I characterize my cash flow inbound up on receipts? How do I build sort of forecast logic doing that based on natural language prompts instead of things that might be a little bit more technical? And that obviously gives them the ability to be able to build scenarios on their own and things like that, too. And I'D say on the payment side of our house, obviously we have a payments hub part of TIS as well. We've introduced a true AI agent that's able to interact with that platform and have plans to sort of expand that out to the cash forecasting side of the platform during 2026. So coming very, very soon as well in a gentic AI offering. Thank you. Tremendous. Thank you, Royston, and thank you for now. John, of course. John available. I think if you want to ask him more questions in the questions tab, he'll be happy to answer through the medium of typing, which is remains available until the end of the session. John. Thank you. I'm sure we'll see. You again in the new year. And we look forward to meeting you again online. So, our next presentation today, another familiar face to Treasury Dragons attendees. Matthias Van Camp is from Cobase. Welcome again, Matthias. Your seven minutes of fame begin now and I'll hand the floor over to you. Thank you, mike. So to get started. I think it's been mentioned before her also in the pre phase of this presentation. I think one of the most important aspects is to be able to get all your data into a centralized platform. So that is all your banking data, so all your balances and your transactions, but also all your other different sources I've seen also in the past, we see a big movement of very much focused forecasts tools that they're now moving to the bank connectivity side of things. But as many of course know, that is not an easy thing. So where does cobase come from? Cobase comes from this pure multibank payment where we essentially connect all your banks and all your different sources for you to this one system. And then from there we are quite easy able to build this additional treasury scopes on top of that. So think of also a cash flow and liquidity forecasting module. So we have done this already for many, many companies all around the globe. I think there are much familiar names in here and let's just go into there. What do you want to achieve? So essentially every finance team wants to have the ability had to view their future cash positions over different entities and different time horizons. And of course you need to consolidate it, for instance, on a business unit or a regional level or overall group level. So the cobase module essentially enables you to generate a rolling forecast plan with also all kind of workflows and approval flows behind it. So you, as a central treasury, are really in control of the forecast. Different legal entities or they've earned business units can all input their own forecast to you. You'll have the ability to accept it, to reject it, or to change it as well. And really this enables you to got all this data to a central point. And of course we connect to your banks, your accounts, but also your ERP system for crucial data such as accounts receivable, accounts payable, maybe you're also operating the FX module within cobase, so you're a fixed settlement will be in there too. You maybe have intercompany loan administration also ministered in cobus, and this really gives you the full suite. Because I also believe the biggest pain is in gathering all the data into this one platform. So what are the abilities that you get out of it? So different output reports are possible. And of course, also cash sheets with historical concurrent cash analysis. So how does it work? We start with defining together with you your different category groups. And categories for the entry now built for you. Every defined group of legal entities or business units need to be defined together and who has access to what data, who can view or enter data, who can manage it, who can modify it, and who can also view the output reports. And of course, with companies all over the world. When the week start? Is it from Monday till Friday or do you follow a different rhythm? Do you want to report on a weekly basis, on a monthly basis, the number of periods and from there we start building your cash forecast. So essentially you'll have the ability to input on a legal entity basis on a weekly or monthly basis, we can enrich and categorize categories because of course we will be receiving from all your banks all the different transactions. And also there we believe there is a strong use case for AI to apply and categorize transactions to actual categories. So all cash flows are submitted to a central point. They are accepted or just. And from their you are able to generate your output reports. And I think the nice thing of it also is that different legal entities all around the globe can also provide their input in their own currency. And for you, on a group level, you can view it from the group base currency. So in essence, you have the ability to create different screens, to create different input, and from there, you will have built essentially this casin cash out suite. So from here, you have the ability to fill in all the details. Of course it's a rolling forecast and also all data that's already readily available in the platform will be pre populated for you. Also making the work way easier for the teams that need to provide this data. So you will have the ability to drill down on basis, weak basis and even on a day basis and really see it from different views. You have the ability here to compare different output inputs and outputs. So for instance, here the actuals and the forecasts and you can field over different periods of time. So for instance, your actuals compared with the forecast and the difference in it, and also forecast plans compared with each other, giving you this full flexibility to run your cash flow forecast within the cobase platform and have it all centralized in this one mean. So of course this is not everything that we do within cobase. There is much more to the platform, but essentially as you now integrate it and sole solutions are the way to go now. So feel free to reach out if you want to explore this for your business. Thanks very much. And good to see the QR code popping up as the standard closing slide in these sessions. Don't leave us, though, even if you do scan the QR code. Stay with us in Treasury. Dragons, we have two more vendors still to go. And Royston, I think I'm going to turn to you first this time for a question for Matthias and Cobase. Hey, Matthias, Great presentation as always. So, you know, people do talk about APIs a lot, and I'm just curious if you can shed some more light and specifically how your Solution is using APIs to provide accuracy or at least richer data for your customers. Yes. So I think APIs where we commonly use it, for instance, also on the very modern ERP, the cloud based ERPs, because it works very good to be able to pull in different data sources from different ERP systems into the platform. But of course, also the reality is that many corporates out there still face this traditional, maybe even on premise ERP systems. So essentially there we offer the full suite with banks we are moving towards APIs more and more. However what we see is that not always all banks can cover all geographies and all functionality globally. So it can be also a mix of APIs host to host EBIX and Swift connections. But of course we will always be looking for the connection that the lift is the most rich flavor of data information and which we of course can see over 1053xml file. And the question on real time data, of course is how real time has it has to be. I think I had a discussion two years ago with Kate, also about it. Near real time or real time, what are you going to do with the data? So if you get intraday statements, for instance, every hour, or maybe even every five minutes or on movement, what is the difference with an API? And how is it going to influence your cash forecast? Excellent. Thank you. Thanks, Matthias. I can see a question from the audience piling in. We'll start with Nadim's. Nadine would like to know how you deal with cash flows deriving from invoice discounting arrangements. Let's imagine we've got a receivables finance factoring program in place. How are you dealing with that kind of thing? Is there a special way of doing it, or how does it work? Debt is always, of course, a part. I think this fair can be very different per customer and also per bank as such. So that's typically something where we set up different workshops for with team members of ours where essentially the client explains their specialties to us and what is special in their location. And then we try to incorporate it as best as possible. However, there could be also some trade offs that they need to manually enter some of the information, so that could be the reality. Nadim, I hope that answers your question. Do feel free to come back to us in chat. If not and Colin would like to know how you deal with it. When data used to build the forecast are duplicated around the ERP or manual forecasting, that duplicate data piece, is that an issue? So essentially your system always checks on duplicates, so as soon as there is a duplicate also on the payments from one of the files will be excluded. However, here it's also that we will try to pre populate data already as much as possible. So if there is an automated connection between the ERP and the manual input, the data will already be filled out with ERP input. So then it's the correction essentially, that the user that does the manual input, they have to change the automated input that is already there. And I think that already prevents a great deal of duplicate information in there. Tremendous. Thank you. If we have no more questions from Dragons, then thank you very much, Matthias. We'll move on to our next presenter and hopefully Mas will see you in the new year. For now, it's time to welcome back to the Treasury Dragon stage, Adib Barak, VP product at Panax. The D, of course. Has been here before. And also at the Working Capital Forum, our event in Amsterdam a couple of weeks back, where cobase was also represented. So, Adi, you know how this works by now. The floor is yours for precisely seven minutes. Take it away. Thank you, Mike. And hi, everyone. Very happy to be here again presenting Panax and our forecast solution. So to give you a brief introduction to Panax. Panax is a trusted AI native cash management solution. We are built for lean teams that have complex treasury needs. And we really see our sweet spot, maybe compared to the people before me, as kind of targeting mid market. Meaning the business is complex enough. There are entities, there are different geographics, different currencies and so on. But the team is still nimble and needs to manage all of it, usually not with too many resources. And this is why we build panics, because we've been hearing a lot of pains from these type of companies on really getting their data, making sure that everything is structured and prepared to be able to build the forecast and to optimize their cash. So when we're looking at the market today and we're seeing what customers are telling us, we are hearing again and again that the forecasting process is still broken and as much of it is very much still manual. Even if some of the companies were speaking with have tried products or tools before, they're usually defaulting to still maintaining the forecast on spreadsheets and in a manual fashion. And why is that so? First of all, it's because the data is fragmented. The data exists in many different places. Banks all over the world, different platforms. Sometimes not just banks. They have investment platforms, payment processors and so on. And then some of the data obviously exists in the erp, not everything that existing. The bank could immediately and very cleanly tell you what this outflow or inflow is about. And getting the data from the erp, combining it with the Bing and creating unified data structure to build the forecast upon with is usually very challenging. And in order to do this, the teams are often relying on other teams within the organization to get the data in, whether it's the IT department or some BI analytics and so on, to allow them to clean the data and prepare it so that they'll be able to use it. So that's another challenge. And then lastly, once they have built it, it needs to constantly, obviously update. And they also need to work in actually gaining insights out of it and not just maintaining it and move on. So at the end of the day, as I mentioned, what we're seeing is a lot of them still defaulting to doing it manually. And not getting the full insight and potential out of this process. So what do these teams need in order to build a straightforward and clean forecast? First of all, unified financial data. Getting to all of these data sources in an automatic way, consistent way. Cleaning the data and building it from the bottom up. Fast and frictionless updates, continuous. Getting the data and continuously cleaning it without any technical bottlenecks. Or dependencies on other teams. And lastly, getting sense out of your data. We are at Panex connecting to all of these platform, cleaning it and allowing you to create a unified cache. Visibility and AI driven forecast. And on top of it, actionable, explainable insights. Not a black box AI, but a way for you to actually understand your data. How do we do it and what does it mean? First of all, the first phase is getting the data, cleaning it up and categorizing it. We are connecting to all of your banks, whether it's VIPIs, SFTP, Swift, or any other method that the bank will allow. And we are using AI to categorize the data based on what we're seeing from the bank, suggesting the category that we see is the best fit, allowing this process to be more automatic and faster. And we're also connecting to the ERP to enrich the data, bringing in data from your customer, suppliers, invoices and so on to suggest a category based on what you already have in your gl. With a reliable data structure and a way for you to really in confidence trust your data, you are then able to start building your forecast. Here. We are also leveraging AI to suggest different patterns in your data based again on your erp. For example, if you're looking at your collection patterns, we are seeing, based on your customer behavior, that maybe your payment terms are set as 30 days, but we are actually seeing from historical data and based on our AI modeling, that the actual days to collect are different than that. And then we will suggest how to place those invoices. Going forward in an automatic way, allowing you to again quickly build up your forecast from the bottom up. We can also allow you to put in any manual data that you are getting from other sources, whether it's a spreadsheet that you share internally with your sales team, whether it's another tool that you're exporting data from, and you are then able to build formulas on top of that. And obviously we also have the ability to use an AI model that will mimic the historical data of the transaction and forecast it going forward. All of that can be built in the structure that you want, whether it's the entity level, category, or subcategory level, weekly or monthly, everything is configurable to match the needs of your business. And then the last part of it is making sense. Out of the data. We are incorporating reports, automatic reports of a rolling forecast. Every week or every month when the forecast is running, we present forecast versus actual report to allow you to see the differences and make changes to the forecast, create additional scenarios and on top of it also using GPT. Sorry, An AI powered insight. An LLM. Allow for some insights on your data that suggest changes. And lastly, we also have an incorporated GPT like experience where you can ask the chat questions and get answers that are tailored again based on your data in a secure cloud that is only based on your data and not shared with external sources. All of this is being achieved relatively quickly for mid market businesses when things are based on API where mostly seeing three to eight weeks of setting it up with a bank connection ERP connection, mapping everything to categories, taking AI, taking advantage of AI to learn and categorize your data automatically and then start building the forecast with you from the bottom up. We also have the ability to input a cache budget and track that and create multiple versions of the forecast to create some scenario planning, also in an automatic way. So to summarize, we are seeing panic adds value to the process for finance team because we are able to get fast time to value powerful and secure AI. We are built for finance executive to achieve release of pain and your expectations and the product works for you. And we could see the impact that we created with other customers here. Sorry, mike. That's okay. No, that was pretty much bang on time. Absolutely. Excellent. Thank you, Adi. So I can see we have questions coming in from the audience, but I'm going to ask Andy Gifford to come up with our first question. If you still have a question that you haven't asked, do type it in now. We still have time. But Andy, can I turn to you for your first question for a D and for panics? Yeah. Thanks, Siddy. Great presentation. I love the way you outlined the problems that we have. And so succinctly. Just thinking about this. You're talking about people coming from Emmanuel spreadsheet system. To a system and looking at your chronologically, you could be up and running within 10 weeks with all of the elements. As a treasurer is thinking about this. Where should they be focusing their preparation to onboard a system like yours you're really concentrating on before they make that giant leap? Yeah, great question. I think it's first to understand what are you trying to achieve from the forecast that will help you understand how to structure your data in a way that fits what you're trying to answer. The easiest way to forecast is to not go into the nitty details of every line items. But that obviously needs to match what you're trying to achieve. So if your goal is to say, I want to know what is the balance that I'm ending my month with, we can help you get there more easily. If you have a clear understanding of the categories that you're trying to forecast, and how do you expect them to behave, whether you want to pull for a specific category the data from the erp, whether you prefer to rely on historical data, or whether you want us to model it using AI. So having a think about the structure and what you're trying to achieve is the best way to prepare as you come to onboard with us. Thanks very much. It's grants. Thank you, Adi. I can see that some other dragons have some questions. Baran, I'll come back to you in just a moment, but let's turn to our audience. Jens would like to know how you handle the fact that there's tons of data that exist in the company but treasury can't access. Everything, but only predefined data. So he says it's a kind of chicken and egg problem. How do you know what data you want if you don't have access to all the data? I can't really see how any kind of TMS can necessarily solve that problem, but any thoughts about that? The problem of think a similar answer to maybe what I mentioned to Andy just now. A lot of it is having the discussion with us and us presenting the options for you and there is a trade off at the end of the day. You either going to try and understand how this team is relying on data and can they provide some sort of an expert for you so that we can pull it into the forecast and for you to create a formula on top of it and just create a pool. But if there's no way to do that, then we can default to the AI modeling and trying to predict based on the behavior of the past. But then it will obviously require us to bring enough historical data to do that. So we have options. We're not magicians, but I think we have enough tools in our toolbox to allow you to solve for that. Excellent. Thank you. I think you had a question. Hi Eddie. Great presentation. At the beginning of your presentation, you mentioned that Penex is an AI native platform. So would you be able to expand on this and explain what it means for corporates in the practical terms? Yes. I think for us, we are thinking about incorporating AI in every layer of our product from the ground up. Whether it's when we are monitoring data issues with the bank, we're using AI to really understand what is the current pattern and when we are encountering issues on the categorization level or cleaning the data. We're using AI to find groups to suggest categories. We are using LLM to map between the ERP and our cash categories. And then the final layer is everything that you're seeing in the product around either the AI modeling of a forecast or the insights that I've just shared about in the chat like experience. So for us, being AI native is really incorporating AI in every layer when we're building our product from the infrastructure all the way up to the user experience. Makes absolute sense. Thank you. Thank you. Adi, Dan, we're not going to get to your question right now, but I invite Addie to go into the question section and perhaps give you an answer directly for now. Addie, thank you very much. We hope we'll see you again next year and Dragons, you'll be back later. But let's move on to our last presenter in our cash forecasting hour, it's the co founder of Palm, Gurjit Panu Gajit. Hello. I believe you're in New York. Is that right? That's right. Temporarily. Excellent. Not that that makes any difference at all to the product, but, hey, it's good to know where people are. So you know how this works. Seven minutes, the floor is yours. Amazing. Thanks, Mike, and pleasure to be back. Before we actually start, let me ask a question to the audience. If resources weren't a problem and it basically had a magic wand and you had the ability to forecast every single category across every single bank account within your organization, would you do it and would you find it to be helpful. I'm curious to hear if you answered yes or even as maybe. I'm excited to share with you what how Palm thinks about forecasting from a truly bottom sub perspective and how that unlocks value for the business. Before that Co CEO Co Founder of Palm, former treasurer with over a decade of experience managing cash for companies like Uber and Levi's and Palm was built from a place of a bit of frustration for me, with a lack of flexibility and granularity that I needed to do my cash flow forecasting. And happy to show you a little bit more about palm. So what is pam? Palm is AI native cash flow forecasting software that sits on top of your TMS or your ERP to enable confidence and accuracy in your cash flow forecasting. So no ripping and replacing of your existing systems. Rather, you're supplementing them with Palm to help enable a more confident and accurate cash flow forecast. We do this by being a place where your data from across your organization comes into one platform, and what you get is a forecast that is quickly generated. For each category within each bank account. It comes with built in various analysis to understand where things might be going off. And granularity that allows for the flexibility of how you report and understand what's happening across your business. But before jumping in, just want to share that palm support multinational companies across the globe, across the US And Europe. We support private companies as well as public listed companies and our industry agnostic. What I'm calling out here on this slide is one of our more recent case studies that we published. In which we supported one of the biggest and fastest growing sportswear brands in the world. On and unlocking $350 million of idle cash from across the globe for them to be able to bring it back to headquarters and include it into their investment program. And it wasn't just about getting there through better accuracy or higher accuracy. It was also around confidence, which is the key with the aim that within the platform, not only do you march towards accuracy, but you also have confidence to understand what's driving or building your forecast, because it doesn't matter how accurate your forecast is if you're not confident or trust in how it was generated, then it's a bit useless in a way, because you don't trust that data. So when building a forecast, we know that not every forecast is created equal or is the same for every company. Each corporate has a different set of access to data and limitations to that data that determine how forecasts are being generated. So what I mean is some companies might have great historical context, but they might lack a view on the near term, maybe AP data. Some might have subsidiaries that provide very sparse forecasts, which I think a lot of us are probably used to, while others have access to maybe company wide data lakes. So it's a myriad of things that come together to create a forecast. And with Palm, depending on what information you have available and how your specific corporate setup is. We work together to build a forecast using the pillars that you see here. So with these pillars, it's a number of things, right? It could be your historical context where we understand what's actually happened across your business. And leverage machine learning models to take that historical activity and extrapolate that forward. That includes seasonality. So if you have a business that is highly seasonal, these models are able to capture that through that historical context and build a forward looking view for that. And then we have our connected data. Right? So that's whether it's your AP AR information that sits in an ERP system. Whether it's your bank statements that's sitting in your tms, you're able to bring that information again into palm so that it starts building these layers of forecasts depending on that information that you have. And then there's user input. Sometimes that's something as simple as an email or a message that comes across to you about some one off transaction that's happening in the future. Or more commonly, when we as treasury teams receive different files from different subsidiaries or business lines and different formats. And how are we, you know, getting that into a platform to then be able to actually utilize. And I'll touch on that in just a second. But that's a user inputs entails. And then finally, there's assumptions, right? We know businesses are not static. They change over time. And we live in a dynamic world. And therefore, in order to really understand where is your cash going to be in the future, We need to be able to add and remove our assumptions and plans that drive that forward looking view. So we can also make sure that that's being taken into account when we're forecasting our cash. And so we have a number of features that address these pillars, but today, with the limited time, all highlight three of them. So the first one we have here is our intelligent upload feature. So this is for teams that are challenged with having to consolidate those multiple spreadsheets or inputs from different teams in subsidiaries and stakeholders into one centralized spreadsheet. They're typically having to copy, paste, format data, get it to a specific structure and then also validate the inputs before actually utilizing the data. Whereas in Palm you can simply just drop the files as they are into the platform, so there's no formatting, no data validation needed to be done by a user. All of that, it's done within the system. WithinPM. You simply load the file. Palm will map the columns to Palm specific categories that drive the cache movements. As always, as a user, you can amend the mapping if it's incorrect, but we rarely see that happen. And then the data is taken from that file and then posted to the bank accounts in the categories that are being driven by it. So essentially you get to a place where you're having a consolidated information that typically could take minutes to maybe even hours to consolidate manually into essential source to here, you're just dropping those files as they are. They're being validated, they're being formatted, and they're being loaded into your cash flow forecast, into your account. And so let's talk about assumptions. So your forecast is as good as your ability to be flexible and nimble enough to change when market conditions or business conditions change. And that's what we focus on in our ability to add these assumptions to your forward looking forecast. So with pom, you bake in these changes you expect in the future, and you could quickly manage and update your assumptions through different dimensions to cover the forward looking views. So whether you want to say it's a group of entities that are increasing revenue in a specific region, or if it's a specific company wide category, you can make those adjustments within Palm layer those and remove those layers as you see fit to complete those assumptions and add into your forward looking forecast. And finally. And finally our variance analysis. We know that in order to have a very proactive variance analysis, the opportunities to go really granular is where you unlock quite a bit of value. So in palm, what we do is we alert you for any variances across any category across all of your accounts. And why this is important is that a lot of cases, treasury teams are doing the variance analysis at a company or group level across balances. But then you have a problem when you have conflicting flows, in which case you might have one subsidiary that's overpaying by a million and one that's received a million more than expected. That nets the balance out to zero. So, as a team, you think your forecast is correct, but in reality, there might be some underlying issues under the surface that you might not have captured. And that's it. I think I'm right on time. If you want to learn more about the product, we do have a webinar that's coming across on January 21st where we go deeper into the product and a view across what we're building for the future as well. We'd love to see you all there. Tremendous. Thank you. And to answer your earlier question, Gurjit, Simon, Jennifer, and Jason all said yes, so got at least three people who absolutely wanted what you were offering. You can see them in chat if you want to reply directly. So for now, Varen, can I come to you, our guest dragon this week and ask you for your first question for. Sure. Thanks, mate. And great presentation, Gurjeet. Good to see the customer success story as well. So the question that I wanted to ask was, in general, there's been chatter within corporates that they can develop their in house solution using AI, as there are a number of startup companies in the market now which are removing this development barrier for them. So the question is, why should companies use your solution as opposed to developing their own in house solution. Yeah, that's a great question. And I think I'd say I'm pretty well placed to answer that. Being formally on the corporate side and now kind of on the vendor technology side, I've learned quite a bit that AI is enabling quite a bit. Right. And it's bringing that barrier to entry. A lot lower to start leveraging AI across your business, whether it's in treasury or anything else. We all hear about Vibe coding, for example. But what gets lost is actually that nuance, that one level deeper, really having AI understand the context of what you're building and getting to those outcomes that we're looking for. It's not really super straightforward, and that's one learning that I've had over the last couple of years, that it does require quite a bit of management training and understanding of how technology speaks to each other to then leverage that kind of outcome. So I think with that, it is definitely doable. For some things. I don't think that just knowing how much time is being spent in building these products. It's not a simple going into a byte coding tool and building an app and having it work for you out of the box. It does require quite a bit of maintenance. Very similar to if you think about what SaaS did to the on Prem stuff. Great answer. Thank you. Thank you for that. I can see we have questions coming in from the audience. I will take another Dragon's question. If anyone has one, do raise your hand, electronically or otherwise. Dan has a question. Gojit, I sort of think you answered this, but Dan's asking, can your system handle diverse file types? Although we don't. Specifically mention PDF. And I don't know what ELS is, but, yeah, let's stick with PDF. And those are kind of files you can analyze. Yeah. So it depends on what that file is being used for. So if it's a way of providing information around maybe bank statements, if you have some statements that are not connected to your TMS and you're receiving a PDF, you can absolutely load that into Palm, which will then parse it and then load it into the platform for you in terms of the forecast. Right now we're supporting CSV's Excels, but then look to expand that out to further over time. Makes absolute sense. Okay, if there are no more immediate questions from the Dragons, I'll say thank you very much, Gertjek. Praise present. Great presentation. And again, we hope to see you next year. In the meantime, we'll move on. And it's time for that difficult moment when we analyze what we've seen again, no winners or losers in Treasury Dragons, but let's just have a quick chat through. And, Andy, if I could start with you did what you saw, match what you hope to see. Yeah, thanks, Mike. And thanks to all the presenters. I was really impressed by the way the proposals are focused on. There's a separate exercise to the normal TMS stuff. And they all recognize the challenges that lean treasury teams have trying to manage this very complex problem of cash forecasting. All the asks that I had around forecasting and scenario planning seem to be very well met. And there were a couple that really clearly demonstrated. They included the ability to make suggestions to treasurers about what they should be doing with the output that they're getting from the predictions. Just a couple of call outs that really impressed me. John Tis I love the focus on collaboration. I think it's something that very often treasurers don't get enough collaboration from other areas in the business. And if you have a tool which promotes that, I think it's really useful. Matthias from CObase using the APIs to gather data is clearly going to make that gathering of data more accurate and much lower to collect. And Addi I love the end game focus of Panex where your flexibility in the pathway. If you're not ready for a big solution, you can start small, which I think is going to stop people being so frightened about taking that leap. And goodjit just a critical nature of having confidence in the forecast and that's really important because very often we go up with a forecast and say, all right, so we'll make this big investment, and we'll say, maybe. I'm not so sure about it. So having that extra layer of confidence that what you're talking about is correct, I think it's really great. I was very impressed by all four of them. I see them moving very quickly. They're adopting latest technology very quickly, and I think there's still more to come, to be honest. Excellent. Thank you. It is interesting to see how much this has changed over the time that we've been doing Treasury Dragons. Kate, were you also impressed with what you saw? Indeed. And I really enjoyed the increasing sophistication or the sophistication use of AI. That's something that I'm focused on right now in my own work, and it was really a pleasure to see. Thank you very much, Kate. Now, Varan, you're a first timer on Treasury Dragons. Did these presenters live up to your expectations? Indeed, Mike and I would say all the presentations were very good and it was great to see the innovations in this space, particularly that all the tech companies and the fintechs are leveraging the latest technologies, whether it's APIs to connect with your banking partners or ERPs, or using AI to for cleansing purposes. Of the data. Because, as we say, data is the new oil for every organization. I believe the focus for the fintech should be on demonstrating factor of their solution. Because the next leap won't come from just how good looking the solution is, but how well the technology is aligned with the real world priorities of the treasury teams. Yeah, it makes absolute sense. Thank you. And final word to you, Royston. Quite often you are the harshest critic of treasury technology vendors. Are you in a generous mood today? Yeah, absolutely. I'm feeling very festive mood today, Mike. In fact, I feel like it's an early Christmas present for us treasures with what I've heard and a couple of snaps sort of sound bites I heard today, which I'm really pleased and impressed with. I think it was tis mentioned a blueprint for managing liquidity. I definitely believe that the big challenges gathering data which co adequately covered. And I like Panics's comment about they have lots of options but they're not magicians. Oh, if only we could use that more often when we are basically the same questions as treasurers. So true, but actually very pragmatic and not least, but Palm saying AI is not perfect, which I've often said it still needs managing. So on those kind of sound bites I would say key challenges still remain. Data quality, which is really up to corporates to continue to try and solve ourselves. Some systems are helping in that respect, and I believe that AI is there to help us as a tool. So we should not discount it, but do not trust it implicitly. And obviously that was stumbling blocks. Where the tools are complex, then user adoption is going to be slow. But the ultimate goal still remain, which is working capital optimization in my view and therefore forecasting or cash flow will not go away anytime soon. And last but not least, let's not forget cybersecurity and Ford risk, we still need to keep an eye on that in the rearview mirror. Or lease, certainly front as well, because that's a huge risk that we still have to manage. Excellent. Thank you very much, Royston. And thank you to all of our Dragons for all your support and help this year. Just as we're about to leave you, we've got one last question to ask of our entire audience. Every year, the Treasury Dragon surveys all of our readers and there are many thousands of us. To ask your feelings about treasury technology and what you plan for the year ahead. It gives us the data for our annual treasury technology report. You now have the chance to take that survey. And I'm just going to fire that link out to all of you now, so feel free to click on the button as we leave. You take the survey. Amazingly, we do offer a small prize for those of you that completed, provided you are bona fide corporate treasurers. And of course, you'll get a copy of the full report slightly ahead of everybody else if you complete the survey for us. So with that, I'll leave you all. Thank you. Very much for being with us throughout this year, and we look forward to next year. In the meantime, have a happy and prosperous festive season and New Year. Thanks again. Bye. Bye. Bye. Bye, everyone.